{"record_type":"pith_number_record","schema_url":"https://pith.science/schemas/pith-number/v1.json","pith_number":"pith:2024:KU7WITXUAOSFI2MPDWNUH5LA3C","short_pith_number":"pith:KU7WITXU","schema_version":"1.0","canonical_sha256":"553f644ef403a454698f1d9b43f560d883b4a64eb921876729326eeeffc4f1df","source":{"kind":"arxiv","id":"2402.05445","version":2},"attestation_state":"computed","paper":{"title":"Accurate LoRA-Finetuning Quantization of LLMs via Information Retention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Haotong Qin, Jie Luo, Michele Magno, Shouda Liu, Xianglong Liu, Xiaoyang Li, Xingyu Zheng, Xudong Ma, Yang Zhang","submitted_at":"2024-02-08T06:53:31Z","abstract_excerpt":"The LoRA-finetuning quantization of LLMs has been extensively studied to obtain accurate yet compact LLMs for deployment on resource-constrained hardware. However, existing methods cause the quantized LLM to severely degrade and even fail to benefit from the finetuning of LoRA. This paper proposes a novel IR-QLoRA for pushing quantized LLMs with LoRA to be highly accurate through information retention. The proposed IR-QLoRA mainly relies on two technologies derived from the perspective of unified information: (1) statistics-based Information Calibration Quantization allows the quantized parame"},"verification_status":{"content_addressed":true,"pith_receipt":true,"author_attested":false,"weak_author_claims":0,"strong_author_claims":0,"externally_anchored":false,"storage_verified":false,"citation_signatures":0,"replication_records":0,"graph_snapshot":true,"references_resolved":false,"formal_links_present":false},"canonical_record":{"source":{"id":"2402.05445","kind":"arxiv","version":2},"metadata":{"license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","primary_cat":"cs.LG","submitted_at":"2024-02-08T06:53:31Z","cross_cats_sorted":["cs.CL"],"title_canon_sha256":"67aa8ac05bbfaf55caa1e1ed007e4b51487f19c41fa98b7fa3d651392a74660e","abstract_canon_sha256":"e091e18e42792a0cf4049bf9de890d763850f79c60977aaec690e75e83a1b2ac"},"schema_version":"1.0"},"receipt":{"kind":"pith_receipt","key_id":"pith-v1-2026-05","algorithm":"ed25519","signed_at":"2026-07-05T08:23:17.748223Z","signature_b64":"QMDYk943EVW5y8oOrGvMvsI+P35aNf1wP1EVTFtAEUiVrr51WrbpWuPHJ90ucXQ2wI/d1MLLW0yjB9GG0PnwBg==","signed_message":"canonical_sha256_bytes","builder_version":"pith-number-builder-2026-05-17-v1","receipt_version":"0.3","canonical_sha256":"553f644ef403a454698f1d9b43f560d883b4a64eb921876729326eeeffc4f1df","last_reissued_at":"2026-07-05T08:23:17.747732Z","signature_status":"signed_v1","first_computed_at":"2026-07-05T08:23:17.747732Z","public_key_fingerprint":"8d4b5ee74e4693bcd1df2446408b0d54"},"graph_snapshot":{"paper":{"title":"Accurate LoRA-Finetuning Quantization of LLMs via Information Retention","license":"http://arxiv.org/licenses/nonexclusive-distrib/1.0/","headline":"","cross_cats":["cs.CL"],"primary_cat":"cs.LG","authors_text":"Haotong Qin, Jie Luo, Michele Magno, Shouda Liu, Xianglong Liu, Xiaoyang Li, Xingyu Zheng, Xudong Ma, Yang Zhang","submitted_at":"2024-02-08T06:53:31Z","abstract_excerpt":"The LoRA-finetuning quantization of LLMs has been extensively studied to obtain accurate yet compact LLMs for deployment on resource-constrained hardware. However, existing methods cause the quantized LLM to severely degrade and even fail to benefit from the finetuning of LoRA. This paper proposes a novel IR-QLoRA for pushing quantized LLMs with LoRA to be highly accurate through information retention. The proposed IR-QLoRA mainly relies on two technologies derived from the perspective of unified information: (1) statistics-based Information Calibration Quantization allows the quantized parame"},"claims":{"count":0,"items":[],"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"source":{"id":"2402.05445","kind":"arxiv","version":2},"verdict":{"id":null,"model_set":{},"created_at":null,"strongest_claim":"","one_line_summary":"","pipeline_version":null,"weakest_assumption":"","pith_extraction_headline":""},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2402.05445/integrity.json","findings":[],"available":true,"detectors_run":[],"snapshot_sha256":"c28c3603d3b5d939e8dc4c7e95fa8dfce3d595e45f758748cecf8e644a296938"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"},"aliases":[{"alias_kind":"arxiv","alias_value":"2402.05445","created_at":"2026-07-05T08:23:17.747792+00:00"},{"alias_kind":"arxiv_version","alias_value":"2402.05445v2","created_at":"2026-07-05T08:23:17.747792+00:00"},{"alias_kind":"doi","alias_value":"10.48550/arxiv.2402.05445","created_at":"2026-07-05T08:23:17.747792+00:00"},{"alias_kind":"pith_short_12","alias_value":"KU7WITXUAOSF","created_at":"2026-07-05T08:23:17.747792+00:00"},{"alias_kind":"pith_short_16","alias_value":"KU7WITXUAOSFI2MP","created_at":"2026-07-05T08:23:17.747792+00:00"},{"alias_kind":"pith_short_8","alias_value":"KU7WITXU","created_at":"2026-07-05T08:23:17.747792+00:00"}],"events":[],"event_summary":{},"paper_claims":[],"inbound_citations":{"count":1,"internal_anchor_count":1,"sample":[{"citing_arxiv_id":"2507.04456","citing_title":"BiVM: Accurate Binarized Neural Network for Efficient Video Matting","ref_index":71,"is_internal_anchor":true}]},"formal_canon":{"evidence_count":0,"sample":[],"anchors":[]},"links":{"html":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C","json":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C.json","graph_json":"https://pith.science/api/pith-number/KU7WITXUAOSFI2MPDWNUH5LA3C/graph.json","events_json":"https://pith.science/api/pith-number/KU7WITXUAOSFI2MPDWNUH5LA3C/events.json","paper":"https://pith.science/paper/KU7WITXU"},"agent_actions":{"view_html":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C","download_json":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C.json","view_paper":"https://pith.science/paper/KU7WITXU","resolve_alias":"https://pith.science/api/pith-number/resolve?arxiv=2402.05445&json=true","fetch_graph":"https://pith.science/api/pith-number/KU7WITXUAOSFI2MPDWNUH5LA3C/graph.json","fetch_events":"https://pith.science/api/pith-number/KU7WITXUAOSFI2MPDWNUH5LA3C/events.json","actions":{"anchor_timestamp":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C/action/timestamp_anchor","attest_storage":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C/action/storage_attestation","attest_author":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C/action/author_attestation","sign_citation":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C/action/citation_signature","submit_replication":"https://pith.science/pith/KU7WITXUAOSFI2MPDWNUH5LA3C/action/replication_record"}},"created_at":"2026-07-05T08:23:17.747792+00:00","updated_at":"2026-07-05T08:23:17.747792+00:00"}